ActionFlow: A Pipelined Action Acceleration for Vision Language Models on Edge
Overview
This paper presents ActionFlow, a system-level inference framework designed to accelerate Vision-Language-Action (VLA) models on resource-constrained edge platforms. The authors introduce a Cross-Request Pipelining strategy that batches memory-bound Decode phases with compute-bound Prefill phases to maximize hardware utilization. Additionally, they propose a Cross-Request State Packed Forward operator and a Unified KV Ring Buffer to fuse fragmented memory operations into efficient dense computations. Experimental results show that ActionFlow achieves a 2.55x improvement in frames per second (FPS) on the OpenVLA-7B model without requiring retraining, enabling real-time dynamic manipulation on edge hardware. The work is available at https://github.com/Frankd35/Action-Flow.
Generated from attributed reports · Updated 2 hours ago
Event evidence and corrections
0 attributed source owners. Ownership does not establish independent confirmation. Quantities are reported separately and are never added together.
Reported quantity · FPS improvement: 2.55 other · Basis not reported
“computations. Experimental results demonstrate that ActionFlow achieves a 2.55x improvement in FPS on the OpenVLA-7B model without retraining, enabling real-time dynamic manipulation on edge hardware. Our work is available at https://github.com/Frankd35/Action-Flow.”
Exact source · revision 1Source owner not reported
Reported quantity · control frequency range: 20 other · Basis not reported
“arXiv:2512.20276v2 Announce Type: replace-cross Abstract: Vision-Language-Action (VLA) models have emerged as a unified paradigm for robotic perception and control, enabling emergent generalization and long-horizon task execution. However, their deployment in dynamic, real-world environments is severely hindered by high inference latency. While smooth robotic interaction requires control frequencies of 20-30 Hz, curr”
Exact source · revision 1Source owner not reported
Reported quantity · VLA model operating frequency: 3 other · Basis not reported
“ent VLA models typically operate at only 3-5 Hz on edge devices due to the memory-bound nature of autoregressive decoding. Existing optimizations often require extensive retraining or compromise model accuracy. To bridge this gap, we introduce ActionFlow, a system-level inference framework tailored for resource-constrained edge platforms. At the core of ActionFlow is a Cross-Request Pipelining strategy, a novel sched”
Exact source · revision 1Source owner not reported
Reported quantity · FPS improvement factor: 2.55 other · Basis not reported
“computations. Experimental results demonstrate that ActionFlow achieves a 2.55x improvement in FPS on the OpenVLA-7B model without retraining, enabling real-time dynamic manipulation on edge hardware. Our work is available at https://github.com/Frankd35/Action-Flow.”
Exact source · revision 1Source owner not reported
Report timeline
Follow attributed reports and material updates.
- arXiv Robotics — research abstractsSignalActionFlow: A Pipelined Action Acceleration for Vision Language Models on Edge
This paper presents ActionFlow, a system-level inference framework designed to accelerate Vision-Language-Action (VLA) models on resource-constrained edge platforms. The authors introduce a Cross-Request Pipelining strategy that batches memory-bound Decode phases with compute-bound Prefill phases to maximize hardware utilization. Additionally, they propose a Cross-Request State Packed Forward operator and a Unified KV Ring Buffer to fuse fragmented memory operations into efficient dense computations. Experimental results show that ActionFlow achieves a 2.55x improvement in frames per second (FPS) on the OpenVLA-7B model without requiring retraining, enabling real-time dynamic manipulation on edge hardware. The work is available at https://github.com/Frankd35/Action-Flow.
Event coverage history
There is not enough continuous observation data to show a trend.
Timezone · UTC
Article dates follow your selected timezone. Briefing editions use Hong Kong time (UTC+8).